In "Black Holes, Information and the String Theory Revolution"
by Leonard Susskind and James Lindesay, the authors give,
(Ch. 15: Entropy of Strings and Black Holes, pg. 170):

"The string and Planck length scales are related by
g^2* (l_s)^D-2 = (l_p)^D-2 (15.0.23) "

They then find the consequences of varying g while keeping
string length l_s fixed, saying, "This implies that the Planck
length varies." [D is dimensionality of the system, g is the
dimensionless string coupling constant, and l_p is the Planck
length.]

Variation of g and l_s seems pretty standard in such studies,
but the implication of variation of l_p, the Planck length, struck
me as a surprising statement in mainstream physics, given
that

l_p = (hbar*G / c^3)^.5

IF the authors are actually suggesting that the Planck length
might vary under certain conditions in our universe, they are
also suggesting changes in one or more of the 'constants' on
the right hand side of the above equation. Unfortunately, they
do not further develop this statement in the book, as far as I
can see.

Is my interpretation correct? Have some authors posited actual
variation in l_p?